CLOSE / RETURN TO INDEX04 / PROJECT
SELECTED PROJECTGenerative AI
Cascaded Text-to-Image Diffusion Model
An Imagen-inspired diffusion cascade separating text-conditioned generation from super-resolution.
1024²OUTPUT RESOLUTION
- Python
- PyTorch
- Diffusion
Objective
Build a text-to-image system that increases output resolution through separate generative stages.
System
Implemented a multi-stage diffusion pipeline inspired by Imagen with Python and PyTorch, separating text-conditioned image generation from super-resolution.
Outcome
The cascade progressively produces images at resolutions up to 1024 × 1024. No formal image-quality score is claimed.